MONETARY POLICY & THE ECONOMY
Quar terly Review of Economic Policy
PO Box 61, 1011 Vienna, Austria www.oenb.at
oenb.info@oenb.at
Phone (+43-1) 40420-6666 Fax (+43-1) 40420-046698
Editorial board Peter Mooslechner, Ernest Gnan, Franz Nauschnigg, Doris Ritzberger-Grünwald, Martin Summer
Managing editor Claudia Kwapil
Editing Karin Fischer, Ingeborg Schuch
Translations Rena Mühldorf, Ingeborg Schuch Layout and typesetting Walter Grosser, Birgit Vogt
Design, printing and production Communications and Publications Division DVR 0031577
© Oesterreichische Nationalbank, 2013. All rights reserved.
May be reproduced for noncommercial, educational and scientific purposes provided that the source is acknowledged.
Printed according to the Austrian Ecolabel guideline for printed matter.
REG.NO. AT- 000311
Visiting Research Program 4
Analyses
Austria Withstands Recession: Return to Positive Growth in Early 2013 6
Klaus Vondra
Structural Budget Balances: Calculation, Problems and Benefits 12
Lukas Reiss
Effective Retirement Age in Austria – A Review of Changes since 2000 29
Alfred Stiglbauer
Event Wrap-Ups
The Future of Sovereign Borrowing
Key Findings of a Conference Jointly Organized by SUERF,
OeNB and BWG on March 8, 2013, in Vienna 52
Ernest Gnan, Johannes Holler
Notes
List of Studies Published in Monetary Policy & the Economy 62
Periodical Publications 64
Addresses 66 Opinions expressed by the authors of studies do not necessarily reflect the official viewpoint of
the Oesterreichische Nationalbank or of the Eurosystem.
(OeNB) invites applications from ex- ternal researchers for participation in a Visiting Research Program established by the OeNB’s Economic Analysis and Research Department. The purpose of this program is to enhance cooperation with members of academic and res earch institutions (preferably post-doc) who work in the fields of macroeconomics, international economics or financial economics and/or with a regional focus on Central, Eastern and South eastern Europe.
The OeNB offers a stimulating and professional research environment in close proximity to the policymaking process. Visiting researchers are expec- ted to collaborate with the OeNB’s research staff on a prespecified topic and to participate actively in the department’s internal seminars and other research activities. They are pro- vided with accommodation on demand and have, as a rule, access to the department’s data and computer re-
Their research output will be published in one of the department’s publication outlets or as an OeNB Working Paper.
Research visits should ideally last between 3 and 6 months, but timing is flexible.
Applications (in English) should include
– a curriculum vitae,
– a research proposal that motivates and clearly describes the envisaged research project,
– an indication of the period envis- aged for the research visit, and – information on previous scientific
work.
Applications for 2013/14 should be e-mailed to
eva.gehringer-wasserbauer@oenb.at by May 1, 2013.
Applicants will be notified of the jury’s decision by mid-June. The following round of applications will close on November 1, 2013.
into recession again. Growth in Austria was driven both by domestic demand and net exports. While the economic outlook for the euro area remains clouded for 2013 as well, Austrian eco- nomic growth is likely to accelerate in the first half of 2013, as forecast by the OeNB. However, the continued weak- ness of order books and uncertainty surrounding the formation of a govern- ment in Italy as well as the latest devel- opments in Cyprus represent potential hurdles for the recovery. HICP infla- tion decreased in January and February after a period of rising and stagnant in- flation rates in the second half of 2012 and is expected to ease further in the course of the year. The economic slug- gishness has had repercussions on the labor market – unemployment is on the rise.
Despite Global Weakness
The first full set of national accounts data for the fourth quarter of 2012 indicates that Austrian GDP shrank by just 0.1% against the third quarter (in real terms, seasonally and working-day adjusted). Thus, output contracted perceptibly less in Austria than in Austria’s main European trading part- ners (Germany: –0.6%; Italy: –0.9%).
Whole-year growth for 2012 amounted to +0.8% in Austria (in real terms, both seasonally and not season- ally adjusted). Growth was fueled not only by net exports, but also by domes- tic demand. The negative contribution to growth of inventory changes corrob- orates the general picture of uncer- tainty among economic agents.
Editorial deadline:
March 25, 2013
1 Oesterreichische Nationalbank, Economic Analysis Division, klaus.vondra@oenb.at. In collaboration with Gerhard Fenz, Friedrich Fritzer, Walpurga Köhler-Töglhofer and Martin Schneider.
Table 1
Results of the National Accounts
GDP Private con-
sumption Govern- ment con- sumption
Gross fixed capital formation
Exports Imports Domestic demand (excluding inventories)
Netexports Changes in
inventories Statistical discrepancy
Change on previous period in % Contributions to GDP growth in percentage points
Q3 11 –0.0 +0.1 +0.1 +1.1 +0.2 +0.5 0.3 –0.1 –0.1 –0.1 Q4 11 +0.2 +0.1 +0.2 +0.7 +0.1 –0.1 0.2 0.1 –0.0 –0.1 Q1 12 +0.4 +0.1 +0.2 +0.2 +0.3 +0.1 0.2 0.1 0.0 0.1 Q2 12 +0.1 +0.1 –0.0 +0.0 +0.7 +0.6 0.0 0.1 –0.1 0.1 Q3 12 +0.1 –0.0 –0.1 +0.1 +0.9 +0.3 –0.0 0.4 –0.2 –0.1 Q4 12 –0.1 –0.1 +0.1 +0.1 –0.3 –0.2 –0.0 –0.1 –0.2 0.2
2009 –3.5 +1.1 +0.9 –6.4 –15.3 –11.8 –0.6 –2.9 –0.6 0.6
2010 +2.2 +1.6 +0.0 +0.7 +8.9 +8.0 1.1 0.8 0.5 –0.1
2011 +2.7 +0.9 –0.4 +6.3 +7.1 +7.0 1.7 0.4 0.5 0.1
2012 +0.8 +0.4 +0.4 +1.8 +1.8 +1.2 0.6 0.4 –0.3 0.0
Source: Austrian Institute of Economic Research (WIFO).
The OeNB’s Economic Indicator Heralds Positive GDP Growth in Austria in Early 2013
The OeNB’s Economic Indicator of March 2013 shows that Austria is re- bounding from the sluggish perfor- mance during the fourth quarter of 2012. Specifically, the OeNB forecasts real GDP to edge up in the first quarter of 2013 (+0.1% against the previous quarter) and to quicken a bit more in the second quarter of 2013 (+0.3%).
This represents an upward revision of 0.1 percentage points in the second quarter of 2013 against the forecast in the OeNB December 2012 Economic Outlook for Austria.
The OeNB’s Economic Indicator of March 2013 reflects the most recent developments of hard and soft facts. For instance, industrial output as defined by Eurostat2 rose astonishingly force- fully in December (month on month) but nonetheless fell slightly for the fourth quarter as a whole (table 2). At the same time, the decline in construc- tion output (NACE Section F) intensi- fied in the course of the fourth quarter.
Hence, industrial output as a whole (NACE Sections B through F3) dimin-
ished in the fourth quarter, notwith- standing a small increase in December.
New orders lead industrial produc- tion by some two to three months, but are also more volatile. The slowdown of the global economy in the fourth quarter of 2012 has left its marks above all on order books. Industry will come to feel the impact of the contraction of new orders during this period with a time lag. Only a marked improvement in actual orders received in the first quarter of 2013 could offset the damp- ening effects of the developments in the fourth quarter.
Confidence indicators reflect the most recent economic developments in a timely manner. At the time of writ- ing, corresponding measures were available up to and including February.
However, the indicators do not present a uniform picture: After having fallen slightly in January, the Economic Senti- ment Indicator (ESI) of the European Commission rose perceptibly in Febru- ary; it has thus moved closer to its long- term average. With the exception of January, the indicator has been improv- ing slowly but surely, signaling an im- provement of sentiment that should
2 Industrial output as defined by Eurostat comprises NACE Sections B through D excluding Group D353, notably manufacturing (Section C).
3 NACE Sections B to F comprises the following production areas: Mining, manufacturing, energy and water supply, and construction.
Table 2
Industrial Output and New Orders (seasonally adjusted)
Dec. 12 Nov. 12 Oct. 12 Q4 12 Q3 12 Q2 12 Q1 12 2012 2011 2010 2009 Change on previous period in %
Industrial output NACE B–F +0.6 –0.5 –0.2 –1.1 +0.9 +2.0 +0.2 +1.9 +5.7 +4.5 –9.5 Construction output –2.1 –1.2 –0.5 –3.0 +2.4 +3.1 –2.3 +1.8 –0.1 –4.3 –1.9 Industrial output as defined
by Eurostat +3.5 –1.1 –0.3 –0.6 +0.3 +1.6 +1.0 +1.8 +6.9 +6.6 –11.4 New orders NACE B–F +0.6 –0.5 –0.2 –1.1 +0.9 +2.0 +0.2 +1.9 +5.7 +4.5 –9.5 New construction orders –2.1 –1.2 –0.5 –3.0 +2.4 +3.1 –2.3 +1.8 –0.1 –4.3 –1.9 New orders as defined
by Eurostat +3.5 –1.1 –0.3 –0.6 +0.3 +1.6 +1.0 +1.8 +6.9 +6.6 –11.4 Source: Eurostat, OeNB calculations.
also translate into a recovery of various economic measures like industrial output in the first half of 2013. Bank Austria’s Purchasing Managers’ Index (BA EMI) dipped in February after having risen somewhat in January, but it remains below the expansion thresh- old. Unlike the ESI, the Purchasing Managers’ Index has been moving side- ways since surging in November 2012 and is not showing a clear trend.
Apart from the latter two indica- tors, the leading indicator of the Aus- trian Institute of Economic Research (WIFO) and the OECD’s Composite
Leading Indicator went through a trend shift in the fourth quarter of 2012. Like the ESI, WIFO’s leading indicator as well as the Ifo Business Climate Index for Germany, which works well as a proxy leading indicator for Austria, have shown a continuous improvement of the economic situation. This corrob- orates the assessment that the rebound will strengthen from the beginning of 2013. However, what will happen after the first quarter of 2013 is still subject to a fair amount of uncertainty. Despite the most recent upward trend, all con- fidence indicators remain below their
Leading Indicators (February 2013)
Chart 1
Economic Sentiment Indicator ESI: New Orders from Abroad Ifo Business Climate Index 120
110 100 90 80 70 60
2008 2009 2010 2011 2012 2013
2008 2009 2010 2011 2012 2013 2008 2009 2010 2011 2012 2013 2008 2009 2010 2011 2012 2013
2008 2009 2010 2011 2012 2013 2008 2009 2010 2011 2012 2013
Source: European Commission.
BA Purchasing Managers’ Index 65
60 55 50 45 40 35 30
Source: Bank Austria.
BA EMI: New Orders 70
65 60 55 50 45 40 35 30 25 20
Source: Bank Austria.
ATX 5,000
4,000
3,000
2,000
1,000
Source: Wiener Börse.
20 10 0 –10 –20 –30 –40 –50 –60 –70 –80
120 115 110 105 100 95 90 85 80
Source: European Commission. Source: Ifo.
long-term average. In addition, the confidence surveys show little signs of a pronounced improvement of order books (chart 1).
The result of the Italian parliamen- tary elections could put a renewed strain on economic developments be- yond the first quarter of 2013. Pro- tracted uncertainty about the political developments in Italy, which is after all the euro area’s third-largest economy, could engender a reassessment of the reform process in Europe in tandem with a fresh loss of confidence on the part of investors. The results of the elections and the events in the subse- quent days already triggered first nega- tive reactions in the financial markets.
Chart 2 shows yields on Italian ten-year government bonds resuming their up- ward movement and a decline in yields on German and Austrian government bonds, apparently a sign of investors’
quest to find safe havens.
The intensification of the debt crisis in Italy in the second half of 2011 had a distinctly negative impact on Austria,
given the close links between the Austrian and the Italian economies.4 The yield spread between Austrian and German government bonds widened noticeably temporarily. Such contagion effects may well occur in the future, too.
The latest developments in Cyprus – staving off national insolvency with a last-minute solution, putting together an EU-ECB-IMF rescue package, involving private savings above EUR 100,000 – will not have a direct impact on the Austrian economy. As the economy of Cyprus is small (0.2% of euro area GDP in 2012), these measures are not expected to have a direct impact on the euro area either. However, the dramatic negotiations within Cyprus and with the EU and the ECB up to late Sunday night on March 24 could have negative repercussions on sentiment in Europe.
HICP Inflation Accelerates Strongly from Mid-2012 and Begins to Ease in Early 2013
After HICP inflation had subsided in early 2012, it resumed its upward
4 Italy is Austria’s second-most important trading partner.
% 8 7 6 5 4 3 2 1 0
Yield Spreads on Ten-Year Government Bonds
Chart 2
Source: Thomson Reuters.
Austria Germany Italy
2007 2008 2009 2010 2011 2012 2013
The crisis of The crisis of confidence in Italy
intensifies
The Austrian banking sector’s The Austrian banking sector’s The Austrian banking sector’s The Austrian banking sector’s exposure to Eastern Europe exposure to Eastern Europe puts Austria under pressure puts Austria under pressure
puts Austria under pressure The problems inThe problems inThe problems in Italy have an impact Italy have an impact
on Austria on Austria
course to reach nearly 3% in fall 2012.
Inflation quickened above all because service prices augmented sharply. Most recently, service sector prices contrib- uted around 1½ percentage points, or nearly half of the total increment, to HICP inflation. The price momentum was driven above all by substantial wage cost increases, robust demand for tourist services and a moderately higher rate of increase of administered service prices.5 Moreover, seasonal factors (in- dustrial goods excluding energy: cloth- ing and footwear) and weather-related factors (unprocessed food: fruit harvest losses) were implicated in the compara- tively strong uptick in inflation in fall 2012. The OeNB expects these devel- opments to be only temporary rather than presaging a revival of inflation.
While inflation results were persis- tently high throughout the fourth quar- ter of 2012, inflation rates in early 2013 are nevertheless significantly lower than the peak rate of autumn 2011. The moderation of inflation by comparison to fall 2011 reflects more moderate de-
velopments above all in the energy sector, but also with regard to processed foods.
HICP inflation abated from 2.9%
from October through December 2012 to 2.8% in January and 2.6% in Febru- ary 2013. Core inflation (HICP exclud- ing energy and unprocessed food) also decreased in February 2013 and stood at 2.5%. Thus, in February 2013, Austrian HICP inflation remained sig- nificantly above the euro area average of 1.8%. The inflation gap between Austria and Germany has widened markedly from August 2012 (+0.1 per- centage points) to February 2013 (+0.8 percentage points). The principal reason for the widening of the inflation gap to Austria’s main trading partner is the faster rise in the price of services and the underlying more pronounced increase of wages in Austria in 2012.
Austrian HICP inflation exceeded the Italian counterpart rate by 0.6 percent- age points in February 2013.
At 2.8% in February 2013 (Decem- ber 2012: +4.1%), the CPI inflation rate of the micro basket – this basket
5 Tuition fees for students who take longer than the standard time to graduate (long-time students) and for students from non-EU countries; medical services; public transportation.
Contributions to growth in percentage points 5.0
4.0 3.0 2.0 1.0 0.0 –1.0 –2.0
HICP Inflation and Contributions of Subcomponents
Chart 3
Source: OeNB, Statistics Austria.
Services (weight: 44.6%) Industrial goods excluding energy (weight: 30.9%) Food (weight: 15.4%) Energy (weight: 9.1%)
Core inflation (HICP excluding energy and unprocessed food, annual change in %) HICP (annual change in %)
Last observation: February 2013
Jan. 07 July 07 Jan. 08 July 08 Jan. 09 July 09 Jan. 10 July 10 Jan. 11 July 11 Jan. 12 July 12 Jan. 13
contains typical daily purchases of food only – was slightly above the HICP in- flation rate. The CPI inflation rate of the basket of weekly goods (containing selected foods, services and fuel) was markedly below the HICP inflation rate (February 2013: 1.6%; December 2012: 3.1%). Annual CPI inflation ran to 2.5% in February 2013, down from 2.8% in December 2012.
The OeNB anticipates a further de- cline in HICP inflation in the course of 2013. Full-year HICP inflation in 2013 is forecast to come to some 2%. This value represents an upward revision from the OeNB December 2012 Eco- nomic Outlook. The key reason for the revision is the acceleration of service prices, as described above.
Unemployment (Eurostat Definition) Climbs to Highest Level since 2009
The lackluster economy in 2012 in- creasingly dampened developments in the labor market. Hence, in the course
of 2012 – especially in the second half – the jobless rate mounted significantly.
In January 2013, the rate of unemploy- ment as defined by Eurostat attained the highest monthly value (4.9%) since the crisis in 2009 (5.2% in September 2009). Youth unemployment (job seek- ers up to the age of 25) augmented to 9.9% in the course of 2012; this rate, though, is still far below the euro area average of 24.2%.
The rise in unemployment was masked by the steady rise in employ- ment in the course of 2012, which itself came as a surprise, given the difficult economic situation. Another special factor that has had an impact on the pace of employment growth was the opening of the Austrian labor market to job seekers from the countries that joined the EU in 2004 (with the excep- tion of Bulgaria and Romania). While these barriers had been removed already in May 2011, nonresidents accounted for some 60% of the incre- ment in employment in 2012.6
6 The calculation refers to the rise in employment from January to December 2012. In December, the share of nonresidents in total employment came to 15%.
Table 3
Labor Market Overview
Payroll employment Unemployment Unemployment rate Vacancies Training 1,000 Annual
change in % 1,000 Annual
change in % AMS
(%) Eurostat
(%) 1,000 Annual
change in % 1,000 Annual change in %
2010 3,360 +0.6 250.8 –3.7 6.9 4.4 31,009 +14.1 73 +14.2
2011 3,422 +1.8 246.7 –1.6 6.7 4.2 32,310 +4.2 63 –13.6
2012 3,465 +1.3 260.6 +5.7 7.0 4.4 29,422 –8.9 67 +5.3
Q1 12 3,403 +1.6 297.3 +4.4 8.0 4.1 27,586 –11.6 68 –1.5 Q2 12 3,462 +1.3 234.3 +5.6 6.3 4.3 32,219 –11.2 68 +6.3 Q3 12 3,537 +1.2 229.9 +6.5 6.1 4.5 31,689 –5.5 60 +6.6 Q4 12 3,460 +0.9 281.1 +6.3 7.5 4.6 26,195 –7.2 70 +10.7 Sep. 12 3,507 +0.7 229.0 +5.0 6.1 4.5 31,609 –3.4 66 +6.4 Oct. 12 3,486 +1.0 249.9 +6.3 6.7 4.5 28,520 –7.7 73 +11.5 Nov. 12 3,466 +0.9 270.4 +6.7 7.2 4.5 25,454 –5.9 74 +9.0 Dec. 12 3,427 +0.7 323.0 +6.0 8.6 4.7 24,610 –7.9 62 +11.9 Jan. 13 3,404 +0.6 338.4 +6.4 9.0 4.9 22,760 –9.6 72 +9.0 Feb. 13 3,415 +0.7 326.4 +5.3 8.7 x 24,757 –9.9 x x Source: Austrian Association of Social Insurance Providers, Austrian Public Employment Service (AMS), Eurostat (Eurostat unemployment rate: seasonally adjusted). Payroll employment
data for February: preliminary data of the Austrian Federal Ministry of Labour, Social Affairs and Consumer Protection.
national legislation have heavily increased the importance of structural balances in economic policymaking in Europe. As defined by the European Commission, structural balances are calculated by subtracting the estimated cyclical component of government revenue and spending as well as certain temporary factors from the headline balance.
Structural balance estimates can be subject to significant measurement errors, which are mainly related to uncertainties about potential output and nonlinear reactions of tax revenue to sharp changes in GDP growth. The definition of temporary factors can also cause substan- tial problems.
While these problems do not render structural balances useless for the implementation of fiscal policy, they imply that policymakers should not aim to reach the target values for the structural balance exactly a specified each year, but rather on average over much longer time periods (unless exceptionality clauses apply). Achieving the targets on average can be ensured by using appropriately specified control accounts.
JEL classification: E62, H6
Keywords: structural balance, cyclically adjusted balance, fiscal rules, fiscal policy
Changes in government revenue and spending are not only driven by discre- tionary policy action but also by changes in economic conditions through the impact of automatic stabilizers. When GDP growth is below trend, tax reve- nue typically grows below trend as well, while social transfers to the un- employed will likely increase. Further- more, changes in spending or revenue can also be driven by noncyclical transi- tory factors like one-off transfers to troubled banks or one-off taxes on wealth or specific forms of fiscal “gim- mickry.” When assessing short-term consolidation needs or long-term fiscal sustainability, one should try to adjust for such factors, i.e. perform the analy- sis on the basis of the structural bal- ance. The (unobservable) structural balance indicates how large the budget balance would have been if the econ- omy were at mid-cycle and if (certain) transitory noncyclical effects had not materialized.
At the European level, the reform of the Stability and Growth Pact (see
for example Holler and Reiss, 2011) and agreement on the fiscal compact (contained in the Treaty on Stability, Coordination and Governance in the Economic and Monetary Union) have heavily increased the importance of structural/cyclically adjusted balances in economic policymaking. Above all, the fiscal compact requires countries to implement rules on structural balances in national legislation (preferably at the constitutional level).
Section 1 of this article describes how cyclically adjusted and structural balances are calculated, essentially out- lining the European Commission’s method. Section 2 discusses the short- comings of these concepts, while sec- tion 3 focuses on how to handle these problems in policy implementation.
1 Calculation of the Structural Budget Balance
According to the European Commis- sion’s method, the structural balance is typically calculated in two steps. First one deducts the cyclical component,
Refereed by:
Geert Langenus, Nationale Bank
van België
1 Oesterreichische Nationalbank, Economic Analysis Division, lukas.reiss@oenb.at. The author thanks Walpurga Köhler-Töglhofer and Geert Langenus for valuable suggestions and discussions.
which is the product of the output gap (section 1.1) and the budgetary semi- elasticity (section 1.2), from the head- line budget balance to arrive at the cyclically adjusted budget balance. Then one deducts certain one-off factors (section 1.3) from the cyclically adjusted balance to arrive at the structural balance.
1.1 Calculation of the Output Gap
The output gap is the relative difference of actual GDP from “potential GDP.”
The latter term can be confusing as it refers to the level of GDP which would prevail if the economy were at the mid- point of the cycle (Larch and Turrini, 2009) rather than indicating GDP at 100% capacity utilization and full em- ployment. Thus, potential GDP should represent a relatively smooth underly- ing trend in GDP, which is estimated by a combination of structural equa- tions and statistical filtering methods.
In the European Commission’s “pro- duction function approach” (D’Auria et al., 2010) GDP is decomposed into labor (L), capital stock (K) and the residual total factor productivity (TFP):
Y=LαK1−αTFP
The trend components of these three elements are calculated as follows:
• The capital stock is calculated by accu- mulating past gross fixed capital for- mation (investment), which is discoun- ted by an annual depreciation rate (perpetual inventory method). The result is then taken to calculate the contribution of capital to potential output; there is no cyclical adjust- ment of the capital stock.
• The labor component is decomposed into the product of working-age population, participation rate, emp- loyment rate and hours worked per person. The developments of the latter three are divided into a trend
and a cyclical component. While the trends of the participation rate and hours worked per person are calcula- ted with an atheoretical HP filter, the trend of the unemployment rate (the nonaccelerating wage rate of unemp- loyment – NAWRU) is computed with a Kalman Filter making use of additional macroeconomic data (wa- ges, terms of trade, …). The product of the working-age population with the trend components of the other elements yields the labor contribu- tion to potential output.
• The TFP component to potential output is calculated by applying a Kalman Filter making use of data on capacity utilization.
Chart 1 shows the European Commis- sion’s estimate of potential growth for Austria from its most recent forecast.
According to the European Commis- sion, potential growth has significantly slowed since the mid-2000s; this is also due to a smoothing of GDP develop- ments around the strong downturn in 2009. The chart also illustrates that whenever actual growth is above (be- low) potential growth, the output gap increases (decreases).
An alternative way to smooth GDP would be to simply use the HP filter on a series of real GDP figures directly.
The HP filter minimizes
t=1
∑
T (yt−τt)2+λ t=2∑
T−1[(τt+1−τt)−(τt−τt−1)]2by choosing an appropriate (unobserv- able) trend τttt and where and where yttt is the loga- is the loga- rithm of real GDP and λ is the so-called smoothing parameter. The European Commission calls the result of this filtering “trend GDP” in its publica- tions. While the HP filter is simpler and easier to replicate than the produc- tion function approach, it is relatively more prone to revisions.
The cyclical adjustment only applies to fluctuations in real variables, i.e.
there is no adjustment for fluctuations in the CPI or any other deflators.
1.2 Defining Cyclical Fiscal Variables and Calculating the Budgetary Semi-Elasticity
The budgetary semi-elasticity indicates by how much the budget balance changes as a ratio to GDP when the output gap increases by 1 percentage point (i.e. when actual GDP increases by 1% for a given potential GDP). The European Commission (Mourre et al., 2013) bases its budgetary semi-elastic- ity measure on the work of the OECD (Girouard and André, 2005).
The first step in calculating the budgetary semi-elasticity is to identify which components of government rev- enue and spending react automatically to the cycle (“passive reaction” of fiscal variables). Discretionary (active) policy measures are by definition not cyclical, even if they come as a direct response to the state of the economy, which can be countercyclical (e.g. stimulus pack- ages as a reaction to a negative output
gap) or procyclical (e.g. lower invest- ment of municipalities in economically bad times due to balanced-budget re- strictions). This distinction between the passive and the active reaction of fiscal policies to changes in cyclical conditions has to be made for two rea- sons:
1. The structural balance should give some information about the amount and direction of discretionary (i.e.
active) fiscal policy.
2. As stated in the introduction, the structural balance should indicate where the headline balance would be if everything were “back to nor- mal.” Including unspecified “typi- cal” active responses of fiscal policy to the cycle in the cyclical compo- nent of the budget balance would make it impossible to assess consoli- dation needs from the size of the structural balance.
On the revenue side it is assumed that all current revenue in taxes and social contributions is cyclical (which make up around 90% of revenue in Austria;
see table 1), while on the expenditure side only unemployment-related pay-
% of GDP 4 3 2 1 0 –1 –2 –3 –4
European Commission Estimate of Potential Growth for Austria
Chart 1
Source: European Commission (autumn forecast 2012).
Labor (persons) Hours per employee
Actual GDP growth Capital accumulation
Output gap TFP
Potential growth
2000 2002 2004 2006 2008 2010 2012 2014
ments are seen as cyclical. Table 2 shows how the European Commission (Mourre et al., 2013) computes the budgetary semi-elasticity of 0.49 for
Austria. Current tax revenue is divided into personal income tax, corporate in- come tax, indirect taxes and social con- tributions; all of them are assigned to a
Table 1
Government Revenue and Expenditure in Austria in 2011
EUR billion % of GDP % of total
Government revenue 144.4 48.0 100.0
Taxes on production and imports 43.1 14.3 29.8
of which: VAT 22.4 7.5 15.5
of which: petroleum tax 4.2 1.4 2.9
of which: employers‘ contribution to Family Burden Equalization Fund 5.0 1.7 3.4 Current taxes on income, wealth, etc. 39.0 13.0 27.0
of which: personal income tax 3.1 1.0 2.2
of which: wage income tax 23.0 7.7 15.9
of which: corporate income tax 5.6 1.9 3.9
Social contributions 48.7 16.2 33.7
Capital taxes 0.1 0.0 0.0
Non-tax revenue 13.5 4.5 9.4
Government expenditure (EDP) 152.0 50.6 100.0
Social transfers 73.9 24.6 48.6
of which: unemployment (COFOG 10.5) 3.6 1.2 2.3
Other current primary expenditure 59.9 19.9 39.4
Interest payments (EDP) 7.8 2.6 5.2
Capital expenditure 10.4 3.4 6.8
Source: Statistics Austria.
Note: COFOG = classification of the functions of government; EDP = excessive deficit procedure.
Table 2
Calculation of the Budgetary Semi-Elasticity for Austria
Share in GDP1 Macro base Elasticity of fiscal variable with regard to base2
Elasticity of base with regard to output gap2
Elasticity of fiscal variable with regard to output gap3
Sensitivity of fiscal variable with regard to output gap
Semi-elasti- city of ratio to GDP with regard to output gap A B C D = B x C E = A x B x C F = A x (D – 1)
Total revenue 0.48 – – – 0.87 0.42 –0.06
Personal income tax 0.11 Wage bill 2.2 0.6 1.31 0.14 Corporate income tax 0.02 Profits 1.0 1.7 1.69 0.04 Social contributions 0.16 Wage bill 1.0 0.6 0.58 0.09 Indirect taxes 0.15 Consumption 1.0 1.0 1.00 0.15
Other revenue 0.04 – – – 0.00 0.00
Total expenditure –0.51 – – – –0.08 0.04 0.55 Unemployment-related expenditure –0.01 Unemployment 1.0 –3.3 –3.30 0.04
Other expenditure –0.49 – – – 0.00 0.00
Budget balance –0.02 – – – 0.47 0.49
Source: Mourre et al. (2013), Girouard and André (2005).
1 Values refer to the average from 2002 to 2011 (see Mourre et al., 2013, for details).
2 Taken from Girouard and André (2005).
3 Taken from Mourre et al. (2013).
Note: The variables in columns C, D and F are called (semi-)elasticities as they could also refer to reactions to changes in actual GDP in % (for a given potential GDP) instead of reactions to changes in the output gap in percentage points.
macroeconomic base variable with which the respective fiscal variable should have a high correlation. Then – based on tax codes – elasticities of the fiscal variables are computed with regard to their respective macro bases; in the case of Austria these elasticities all cor- respond to 1, except for the one of the progressive personal income tax. To ar- rive at the measure of sensitivity with regard to the output gap, these vari- ables are then combined with estimated elasticities of the macro variables with regard to the output gap and the share of the fiscal variable in GDP. Finally, the budgetary semi-elasticity is com- puted by transforming these sensivities into the semi-elasticities of the revenue and expenditure ratio with regard to the output gap and adding them up.2
1.3 From Cyclically Adjusted to Structural Balances: Adjustment for One-Off Effects
To assess the underlying budgetary position of a country, it is also seen as necessary to adjust for (certain) one-off effects. The updated code of conduct published by the European Commis- sion (2012, page 4) on the Stability and Growth Pact and on stability and con- vergence programmes gives the follow- ing general definition: “One-off and temporary measures are measures hav- ing a transitory budgetary effect that does not lead to a sustained change in the intertemporal budgetary position.”
The European Commission (2006, page 114) provides more detail and also gives a relatively long list of examples.3 More recently, large capital transfers to banks in the context of government in- terventions due to the financial crisis have also been (at least partly) recog- nized as temporary measures. Notable is an asymmetry between deficit-in- creasing and deficit-decreasing mea- sures, the latter being more likely to be deducted for computing the structural balance (for reasons of prudence). For example, temporary tax cuts are not accounted by the European Commis- sion as temporary measures, while temporary increases are.
1.4 Example: Structural and Cyclical Budget Developments in Austria since 2003
Chart 2 shows how Austria’s structural balance has evolved since 2003. Due to consolidation measures in the early
2 Note that column F in table 2 refers to the number currently employed by the European Commission (Mourre et al., 2013), while the sensitivity of the budget balance in column E is based on the previous method (European Commission, 2005).
3 Tax amnesties implying a one-off tax payment, sales of nonfinancial assets, exceptional revenues linked to the transfer of pension obligations, changes in revenues or expenditure owing to court or other authorities’ rulings, exceptional revenues from state-owned companies, short-term emergency costs associated with major disasters or other exceptional events, securitization operations and temporary legislative changes in the timing of expenditure or revenues (the latter two only when they have a positive impact on the budget balance).
% of GDP 2 1 0 –1 –2 –3 –4 –5
Structural, Cyclically Adjusted and Headline Budget Balance in Austria
Chart 2
Source: European Commission (autumn forecast 2012).
Structural balance One-off effects Cyclical component Headline balance
Cyclically adjusted balance
2003 2004 2005 2006 2007 2008 2009 2010 2011
2000s, the structural balance was above –1% of GDP in 2003 and 2004, and the contribution of the cycle was negative due to low growth and in 2004 there was moreover a negative one-off effect.4 Then the structural balance worsened significantly in 2005 and 2006 mainly due to cuts in income tax rates. At the same time cyclical condi- tions improved, which led to a decrease of the headline budget deficit to about 1% of GDP in 2007 and 2008. The strong downturn in 2009 was accom- panied by a deterioration of the head- line deficit by more than 3 percentage points, which according to the Euro- pean Commission’s method was mainly due to cyclical factors (the output gap worsened by almost 5 percentage points;
see chart 1), with structural factors (like the cut in the personal income tax and several smaller stimulus measures) playing a secondary role. After a slight structural deterioration in 2010 there was a strong improvement in the head- line deficit in 2011, which was partly
due to a return of the cyclical compo- nent to close to zero and to several con- solidation measures.
1.5 Example: Subnational Cyclical Components in Austria
The recently implemented rule on the structural balance in the Austrian Stability Pact applies to all levels of government. Therefore it does not suf- fice to consider estimates of the cyclical component of the general and the fed- eral/central government balance; cor- responding estimates also need be per- formed for states and municipalities. In Austria, this is simplified by the fact that taxes are primarily collected by the federal government (and then shared with states and municipalities) and that the federal government is also responsible for unemployment insur- ance. Therefore, regional business cycles play only a very limited role for the cyclicality of budget balances.
Allocating the fiscal variables in table 2 either to the federal government (in-
% of GDP 6 4 2 0 –2 –4 –6 –8 –10
Changes in Cyclically Adjusted Balances of 2007
Chart 3
Source: European Commission.
Real time = European Commission spring forecast 2008. Recent = European Commission autumn forecast 2012.
CAB in real time Revision in headline balance CAB in recent forecast Revision in cyclical component
BE DE IE GR ES FR IT LU NL AT PT FI
4 The figure starts only in 2003 as the European Commission does not report one-off effects for previous years.
cluding social security funds) or to the states and municipalities produces a rough picture of the composition of the cyclical component under the Austrian Stability Pact: the federal government (including social security funds) ac- counts for seven-ninths of the cyclical component and the states and munici- palities for two-ninths.
2 Problems in Estimating Structural Balances
Charts 3 and 4 indicate two major problems of the concept of cyclically adjusted budget balances:
1. Estimates of the cyclically adjusted balance are subject to sizeable revi- sions, as is evident from the range of measures published for the pre- crisis year 2007 for the original 12 euro area countries (chart 3). While some of the changes are due to revi- sions of the headline budget bal- ances (yellow bars), the major
changes relate to revisions of the cyclical component (red bars).
These revisions have been above 1%
of GDP in most euro area-12 coun- tries.5
2. Cyclically adjusted balances can worsen significantly within rela- tively short periods of time without the implementation of much fiscal stimulus or consolidation (see chart 4 with figures for Ireland). After having originally estimated Ireland to report a cyclically adjusted bal- ance of roughly 0 in 2007, the Euro- pean Commission has since revised this figure to –1.5% of GDP and it even assumes this figure to have deteriorated to –10% of GDP until 2009. This is somehow counterin- tuitive as there were some tax in- creases in this time span and as pri- mary expenditure growth was much lower than before 2007.
% of GDP 4 2 0 –2 –4 –6 –8 –10
Cyclically Adjusted Budget Balance of Ireland
Chart 4
Source: OeNB, Eurostat, European Commission.
Real time = European Commission spring forecast t+1 (adjusted for later revisions of headline balance) 11/2012 = European Commission autumn forecast 2012 (adjusted for impact of support to banks)
CAB in real time (adjusted) CAB as of 11/2012 (adjusted)
2003 2004 2005 2006 2007 2008 2009 2010
5 Hughes Hallett et al. (2012) analyze revisions in output gap and CAB estimates for OECD countries since the mid-1990s. They do not directly show revisions in cyclical components, but these components can be (roughly) estimated by multiplying the revisions in the output gap with the respective budgetary sensitivity. Assuming a ary sensitivity. Assuming a ary budgetary sensitivity of 0.44 (the OECD average in Girouard and André, 2005), this would yield a mean RMSE ary sensitivity of 0.44 (the OECD average in Girouard and André, 2005), this would yield a mean RMSE ary (root mean squared error) of somewhat more than ½% of GDP when comparing the estimate from t+1 to later ex post data.
2.1 Uncertainty about and Volatility of Potential Output
Typically the largest problem in esti- mating structural balances is the esti- mation of the output gap, which is an unobservable variable. These uncer- tainties can lead to substantial ex post revisions of output gaps and implausibly large swings in potential growth rates.
Both problems are exemplified by chart
5 for Ireland: The European Commis- sion’s estimates and projections of potential growth were revised down- ward substantially from May 2006 to November 2012. Lower potential growth in past years implies relatively higher output gaps and therefore ex post downward revisions of structural balances (chart 3). Furthermore, down- ward revisions of potential growth
Annual change in % 7
6 5 4 3 2 1 0 –1 –2
Estimates of Potential GDP in Ireland
Chart 5
Source: European Commission.
As of 5/2006 As of 11/2012
2003 2004 2005 2006 2007 2008 2009 2010
Table 3
Potential Output Estimates of European Commission and OECD
Output gap Average potential growth NAWRU / NAIRU 2007 2011 2000–2007 2007–2014 2007 2014
EC OECD EC OECD EC OECD EC OECD EC OECD EC OECD BE 2.4 2.8 –0.2 –0.1 1.9 1.8 1.1 1.4 7.7 8.0 7.5 7.9 DE 2.1 2.4 0.3 –0.5 1.3 1.3 1.3 1.5 8.6 8.0 5.1 6.7 IE 3.6 8.5 –2.8 –7.6 5.4 5.3 –0.1 1.7 5.8 7.6 15.2 10.6
GR 3.4 7.4 –10.0 –9.0 3.8 3.0 –1.4 –0.1 10.2 9.9 16.2 12.3
ES 2.1 2.7 –4.2 –6.1 3.5 3.4 –0.0 1.4 11.9 12.6 25.7 16.5 FR 2.9 3.2 –1.6 –2.0 1.8 1.7 1.1 1.4 8.8 8.5 10.3 9.0 IT 3.1 3.3 –1.8 –2.8 1.1 1.2 –0.1 0.4 7.3 7.4 10.5 7.6 NL 2.3 3.3 –1.8 –1.0 2.0 1.9 1.0 1.2 3.5 3.8 5.0 3.8 AT 2.3 3.7 –0.1 –1.3 2.2 2.1 1.2 1.8 4.3 4.3 4.2 4.3 PT 0.9 1.4 –2.6 –3.5 1.5 1.6 –0.3 0.4 9.1 8.1 15.1 11.0 FI 5.1 6.7 –1.5 –0.4 3.0 2.6 0.8 1.2 7.0 8.2 7.4 8.4 Source: European Commission, OECD (autumn forecasts 2012).
forecasts mean that (for a given legisla- tion on revenue) lower real expendi- ture growth is necessary to keep the structural balance constant. Taking the recent estimate of the European Com- mission for Ireland at face value (and again assuming no structural changes on the revenue side), from 2000 to 2007 an average real expenditure growth of 5.4% per year would have been sufficient to prevent the structural balance from deteriorating, while mea- sures to keep real expenditure constant would lead to a slight worsening over 2007 to 2014 (table 3). While it cannot be neglected that potential growth can and does change over time (due to innovations, structural reforms, de- mographic changes, …), variations in potential growth rates of such a magni- tude make it difficult to interpret the levels of, and the changes in, structural balances.
The uncertainty about potential output is also indicated by table 3, which gives an overview of the Euro- pean Commission’s and the OECD’s recent estimates of output gaps for 2007 and 2011 and average potential
pre- and post-crisis growth rates in the 11 largest euro area economies. It shows that potential growth is esti- mated to have declined substantially over time in several countries, espe- cially in Ireland, Greece, Spain and Portugal. Furthermore, the estimates of the European Commission and the OECD differ remarkably. For example, while projecting similar unemployment rates for Spain in 2014, the estimates for the nonaccelerating inflation rate of unemployment differ by almost 10 per- centage points for that year.6
2.2 Nonlinear Reactions of Tax Revenue to Sharp Changes in GDP Growth
The previous assumption of no struc- tural changes on the revenue side (i.e.
of “standard” reactions of tax revenue components to changes in GDP) is not as innocent as it may sound. While the revenue elasticities used by the Euro- pean Commission (Mourre et al., 2013) would imply that tax ratios are roughly constant over the business cycle with- out any policy measures, the experi- ence of the last years has shown that de-
6 Another issue in this context is raised by Kempkes (2012), who shows that there is a negative real-time bias in the estimation of output gaps (and therefore cyclical components) by international institutions, implying a systematic overestimation of structural balances in real time.
% of GDP 2.5 2.0 1.5 1.0 0.5 0.0
Revenue from Taxes on Transactions in Property and Financial Assets1
Chart 6
Source: Eurostat.
1 Stamp taxes (D.214B) and taxes on financial and capital transactions (D.214C)
IE ES AT
1995 1997 1999 2001 2003 2005 2007 2009 2011
pending on several factors (which will be explained below) tax revenue can vary much more or much less than GDP. For example, while from 2007 to 2009 the tax ratio plummeted in coun- tries like Ireland and Spain, it actually rose in Austria despite a sizeable cut in personal income tax rates (with some minor tax increases and decreases roughly cancelling each other out).
One of the main reasons for this development was the asset price bubble in Ireland and Spain. Chart 5 shows the revenue from taxes on property and financial transactions in Ireland, Spain and Austria (in all three countries this revenue is coming primarily from prop- erty transactions). These are indirect taxes and therefore the GDP ratio should – given the elasticity of 1 used by the European Commission – remain constant over the cycle (unless there are changes in tax rates and/or bases).
However, this was obviously not the case: While tax revenues remained sta- ble at roughly ¼% of GDP in Austria, they increased in Ireland and Spain from 1% of GDP in 2001 to 2% of GDP in 2006, only to decrease to less than 1% of GDP in 2008. Due to the elastic- ity of 1 both these changes were identi- fied as structural. The effect of the build-up and burst of the property bub- bles in these two countries becomes also visible when looking at other taxes related to immovable property (like VAT and capital gains taxes). For exam- ple, in Ireland taxes on capital gains made up 1.6% of GDP in 2007 and 0.3% of GDP in 2009 (amidst a de- crease in nominal GDP of around 15%).
Two less severe examples for non- linear reactions of tax revenue to sharp cyclical changes could be observed in
Austria in 2009, when GDP decreased by 3.5% in real terms. That year reve- nue from corporate income tax (which made up 2.2% of GDP in 2008) decreased by 34%, while according to the elasticities employed by the Euro- pean Commission it should have only decreased by roughly 6%.7 At the same time, compensation of employees and therefore wage-related taxes and social contributions developed much better than what would have been predicted by the semi-elasticity of the European Commission. However, these two non- linear effects roughly cancelled out in 2009.
2.3 Some Crudeness in Assessing the Cyclicality of Tax Revenue
At least in theory, structural balances play an important role for fiscal gover- nance in Europe. Therefore, a harmo- nized treatment of different countries can be argued to be very important.
However, this can come at the cost of accuracy. Three examples will be pro- vided in the following, namely the elas- ticity of personal income tax, the prob- lem of noncyclical tax revenue and the elasticities used for indirect taxes.
Girouard and André (2005) them- selves point out that the elasticity of the (typically) progressive personal income tax might be overestimated as they as- sume in their calculation of the tax elasticity that all fluctuations in the wage bill are in wages per person (and that there are none in employment).
For example, when assuming that half of the fluctuations in the Austrian wage bill are driven by fluctuations in em- ployment in persons, the budgetary semi-elasticity would be overestimated by roughly 0.04.
7 There were no major changes in the rate or base of corporate income tax in 2009; the only stimulus measure affecting corporate income tax was a temporary acceleration of depreciation which was projected by the govern- ment to dampen revenue only from 2010 on.
